Subject: some success, some troubles
To: None <port-pmax@sun-lamp.cs.berkeley.edu>
From: Andrew Gallatin <gallatin@isds.Duke.EDU>
List: port-pmax
Date: 06/25/1994 22:05:20
Hi,

Let me prefix this by saying that since sun-lamp never seems to be up
anymore, I've grabbed the ksrc from ftp.iastate.edu.  I realize these
may be out of date, so if what I mumble about next is now fixed,
please ignore it.  (Oh, could somebody send me the file(s) that sup
uses to determine what files to grab.  This way when sun-lamp comes
back up, I can trick sup into just updating stuff, instead of grabbing
everything..).

My first problem was that it wasn't finding any swap space.

In order to get it to boot, I've managed to get the swap added via
some hacks.  In the file src/sys/arch/pmax/dev/rz.c I uncommented some
code in rzgetinfo.  It doesn't seem to read ultrix disk labels, so
changed the default disk label to exactly match the partition info on
my root disk.  This seemed to get the swap added.  At least there are
no more complaints ;-)

My next problem deals with memory allocation.  It boots, (ultrix) init
starts,  and starts (ultrix) sh.  After the execve of /bin/sh, a break
call is issued, which fails with ENOMEM, sh dies, and init tries again
to create sh, and so on..

I've traced this back to vm_map_insert() (in src/sys/vm/vm_map.c)
failing, due to the call to vm_map_lookup_entry() returning true, which
I guess means that the requested address is part of an existing entry.
The call to vm_map_insert looks like:

map=0xc0416400, object=0x0, offset=0x0, start=0x10001000, end=0x10002000

Has this bee seen before/fixed?

Drew


------------------------------------------------------------------------------